High Level Monographs- why?

I recently got two checks in the mail: (1) $500.00 honorarium for a talk I gave at a University (more than I thought it would be), and (2) $11.00 for book royalties for Bounded Queries in Recursion Theory. Perhaps I should talk more more and write less. I talk much faster than I write, so I could really rake it in.

Why do we write high level monographs that very few people will buy? Should we?

1. We are delusional. We think that a book will sell and make us real money. (I never thought this for my book.)
2. We want to get a certain body of knowledge out there. (Yes for my book, though I later wrote a survey gems.pdf, gems.ps. that did a much better job. This is partially because AFTER co-writing the book (co-author Georgia Martin) I knew what I wanted to say.
3. We want an excuse to learn a field. (Yes for my book, and even more so for a book I am working on on van der Warden stuff. See later in this post.)
4. We write books to help us get promotions. In terms of time spend, papers are much better for Tenure. For Full Prof books may be okay. (This is not why I wrote my book, though I think it helped my Full Prof case.)
5. We are intrigued by the mathematics that dicates that the book cost $80.00 for you to buy, and for each copy my co-author and I split $5.00.
6. We like the fact that if there is a mistake it's hard to correct, and once a new result is discovered its hard to insert.

Why do we go through a publisher? Note that our goals and a publishes are different. If I found out that there were illegal copies of my book in China I would be delighted!. And surprised. My publisher would not be delighted, though they may be surprised. My goal is to get the information out there. I do not care about the money (this is not altruistic--- we are talking about $11.00). Also, we can update much more easily if all is online. So why do we use publishers? They lend a certain credibility that chairman, deans, and even our colleagues recognize and respect. We need a way to certify that book is valid in some form without going through a publisher. If someone knows of such a way already in progress, please post a comment. This would be a boon to the community and should not be that hard. At least, it seems easier than the Journal problem.

Having said all this, there are two advantages to having a publisher

1. If people refer to a particular theorem or page in the book, its bad if the book keeps changing. I don't take this seriously since the book won't change THAT much and this should not be much of a problem. Of course, you don't quite need a publisher for this, you just need discipline to not change stuff.
2. The books may never get finished. I have 160 or so pages of a book on VDW stuff (co-authored with brilliant undegraduate Andy Parrish) that is on my website. (I am not supplying a pointer- I want to polish it some more before advertising it.) I was planning on getting it into a reasonable state and then blogging about it. But I keep wanting to add more. And its never quite finished. And I don't have a publisher telling me ``The draft is due on Nov 1, 2007'' . If I did then I would be forced to find a reasonable stop point.